I’ve spent most of my career being pretty skeptical of egalitarian arguments, and in many ways, I remain skeptical. Common though it may be among my philosopher colleagues, the idea that justice requires any robust form of material equality always struck me as both philosophically groundless and economically naïve.
My anti-egalitarian leanings have strong philosophical support. Robert Nozick’s Wilt Chamberlain example, for instance, makes a convincing case that equal outcomes are in deep tension with respect for individual liberty. The leveling down objection (in either its philosophical, literary, or musical form) demonstrates that making a society more equal doesn’t necessarily make it better in any respect. And, perhaps most importantly, Harry Frankfurt showed that what really drives many of our objections to unequal distributions is actually a commitment to sufficiency—the belief that people should have enough, not necessarily the same.
But all of that is perfectly compatible with the claim that less inequality, on the margin, might turn out to be very important after all. That case starts with Adam Smith—not the cartoon Smith of introductory economics, but the Smith who spent hundreds of pages in The Wealth of Nations documenting how the wealthy rig the rules. Smith saw with perfect clarity that employers don’t just compete in markets. They also collude. “Masters are always and every where in a sort of tacit, but constant and uniform combination, not to raise the wages of labour above their actual rate.” And when the legislature steps in to adjudicate? “Its counsellors are always the masters.”
The philosopher Eric Schliesser, in his book on Adam Smith, argues that Smith’s “theoretical partiality toward the working poor” was not sentimentality. It was a deliberate correction for a bias Smith regarded as structural and permanent. “Intentionally favoring the working poor may, in fact, produce more equal treatment of all,” Schliesser writes, because “in most circumstances the rich can take better advantage of any system of rules.”
Throughout the Wealth of Nations, Smith “calls attention to what contemporary economists call ‘rent-seeking behavior’ by well-connected elites.” On Smith’s telling, the entire mercantilist system was a monument to it. The rich didn’t just happen to benefit from the rules. They wrote them.
Later thinkers have formalized the mechanism Smith described. Michael Munger and Mario Villarreal-Diaz argue that the transition from capitalism to crony capitalism has the structure of a prisoner’s dilemma. In any successful market economy, there comes a point where, for a given firm, “it becomes more profitable, in an accounting sense at least, to use the power of the state to extract resources from others or to protect those existing products from competition.” When that point arrives, rational firms redirect resources from innovation to lobbying. Brink Lindsey and Steven Teles, in The Captured Economy, document what this looks like across finance, intellectual property, occupational licensing, and land use: “the proliferation of regressive regulations that redistribute wealth and income up the economic scale while stifling entrepreneurship and innovation.” And a more recent study of corporate mergers finds that consolidation is followed by significant and persistent increases in political influence activities, providing empirical verification of Henry Simons’s basic intuition: economic concentration doesn’t stay in its lane. It spills into politics.
***
Now, if you’re a classical liberal reading this, you probably have a ready response. I know, because I’ve made it many times myself. The response goes like this: the problem isn’t markets, it’s the state. If government didn’t have the power to hand out favors, there would be no favors to seek. The solution to cronyism is to shrink the state.
There is something to this. The classical liberal tradition has always understood that a state with discretionary authority over economic life creates opportunities for rent-seeking, and that one way to reduce rent-seeking is to reduce the authority available to be captured. But while this is the natural first move in the classical liberal response to concentrated wealth, it doesn’t solve the problem. “Just shrink the state” can mean many things; it could mean to spend less, to hire fewer people, to enforce the laws less strictly, or to deliberately scale back capacity in other ways. Unfortunately, some of those ways can be counterproductive or even dangerous. To see why, I want to turn to what is arguably one of the most important papers in political economy of the last two decades: Daron Acemoglu and James Robinson’s “Persistence of Power, Elites, and Institutions.”
Acemoglu and Robinson distinguish between two kinds of political power. De jure political power is the formal, on-paper power allocated by laws and constitutions—the power to vote, to hold office, to write legislation. De facto political power is the informal, real-world power to shape outcomes outside the formal channels—through lobbying, campaign finance, control of media, the revolving door between industry and regulatory agencies, and in some cases outright intimidation or violence. De jure and de facto power interact, but they are not the same, and they are not distributed in the same way. A society can change who has de jure power without changing who has de facto power. And, crucially, those who lose de jure power through institutional reform have strong incentives to invest in de facto power to offset the change. The smaller and wealthier the elite, the more they stand to gain from controlling politics, and the easier it is for them to solve the collective action problem involved in doing so.
Acemoglu and Robinson model the interaction formally and derive a striking result. Under certain conditions—conditions that turn out to be fairly common empirically—the effect of a change in formal political institutions is completely offset by changes in investment in de facto power. This is what they call “captured democracy”: a pattern in which democratic institutions survive on paper but “choose economic institutions favoring an elite.” Even more strikingly, their model shows that in some configurations, pro-elite economic outcomes are actually more likely under democracy than under autocracy, precisely because the elite invests so heavily in de facto power to offset the formal democratic advantage of ordinary citizens.
The historical example that motivates their paper is the American South after the Civil War. Slavery was abolished. Former slaves were enfranchised. On paper, the de jure distribution of political power had been radically transformed. But the southern economic elite, confronted with this formal change, invested massively in de facto power: monopsonistic labor arrangements, policies designed to impede labor mobility, political disenfranchisement through literacy tests and poll taxes, and the systematic use of intimidation and violence to enforce a political order that formal law no longer authorized. The result was that the pre–Civil War economic order—plantation agriculture, labor repression, concentrated wealth—persisted for nearly another century. The formal institutions had changed. The distribution of de facto power had not. And it was de facto power that determined outcomes.
The implication for the “just shrink the state” strategy is sobering. You can strip government of its power to hand out favors, but the concentrated wealth that benefited from those favors remains intact, and the people who hold it have every incentive and resource to rebuild the state in self-serving forms. The state doesn’t stay shrunk. It grows back in the shape the most powerful interests prefer. The “just shrink the state” approach treats political institutions as a dial you can set; Acemoglu and Robinson show that the setting of the dial is endogenous—continuously pushed by the distribution of economic power itself.
***
None of this, I want to be clear, refutes the classical liberal tradition. It does refute the crudest version of the “just shrink the state” response—but that’s only the entry point to a much richer body of classical liberal thinking about institutional design. The tradition has been grappling with variations of the de jure / de facto problem for a long time, and it has developed some genuinely valuable resources for thinking about it. Let me mention two.
The first comes from the public choice economist James Buchanan, especially in his late work with Roger Congleton, Politics by Principle, Not Interest. Buchanan’s starting point was a worry very close to the one Acemoglu and Robinson formalize: ordinary democratic politics tends to produce legislation that benefits concentrated interest groups at the expense of diffuse majorities, precisely because the concentrated interests have more to gain from political investment. His response was to propose what he called the generality principle: a constitutional-level constraint requiring that political decisions apply non-discriminatorily to everyone. If legislation must treat all citizens the same—no targeted subsidies, no industry-specific tax breaks, no carve-outs for favored groups—then the payoff to lobbying collapses. There are no group-specific rents to capture, and therefore no reason to invest in the de facto political power that Acemoglu and Robinson worry about. F. A. Hayek had long championed the virtues of general, abstract rules on the grounds that they preserve individual freedom and accommodate dispersed knowledge; Buchanan’s contribution was to show why such rules are also, specifically, capture-resistant. The generality principle won’t solve every problem, and enforcing genuine generality is itself politically difficult. But it identifies a structural feature of institutional design that makes the de facto power problem less severe—and it does so in a way that is entirely consonant with classical liberal commitments to the rule of law.
The second resource comes from the Nobel laureate Elinor Ostrom, whose lifelong work on commons governance offers something even more radical. The debate about concentrated economic power has typically been framed as a binary: either the state constrains it through centralized regulation, or the market solves it through competitive pressure. Ostrom spent her career showing that this binary is false. Communities can and do self-organize complex governance arrangements—for fisheries, forests, irrigation systems, groundwater basins—that neither rely on centralized state authority nor reduce to pure private ownership. The institutional form she documented and analyzed is called polycentric governance: decisionmaking distributed across multiple overlapping centers of authority, each with its own rules, each accountable to a different constituency, each capable of interacting with and constraining the others.
Polycentric governance has several attractive features. It is often more adaptive than centralized systems, it makes better use of local knowledge, and it is more responsive to the actual preferences of the people it governs. But the feature most relevant to our present concern is one Ostrom called redundancy. In a polycentric system, governance isn’t concentrated in a single institution that determines outcomes for everyone. It’s distributed across what Ostrom referred to as “parallel autonomous systems”—multiple bodies of rulemaking, interpretation, and enforcement operating alongside one another rather than answering to a single central authority. And that distribution has a crucial property: if any single node in the system is captured or fails, the others continue to function. “The probability of failure throughout a large region,” Ostrom wrote, “is greatly reduced by the establishment of parallel systems.”
This is exactly the structural response that Acemoglu and Robinson’s analysis calls for. The de jure / de facto problem is fundamentally a problem of single-point capture: concentrated wealth seeks out the lever of formal authority and bends it to its purposes. A polycentric system denies concentrated wealth that lever, because there is no single lever to find. Capturing one institution leaves several others standing. And unlike Buchanan’s top-down constitutional approach, which depends on getting the rules right at a founding moment and then holding them in place against pressure, Ostrom’s polycentric systems are continuously remade from below by the people who live within them. When one arrangement fails, participants modify it. When capture starts to happen in one forum, others provide alternatives. Governance becomes adaptive rather than fixed, which makes it that much harder to seize permanently.
There’s a further implication of the polycentric logic worth dwelling on, because it cuts against the grain of much progressive thinking about inequality. In a genuinely polycentric system, wealth inequality itself can sometimes be a source of strength. Multiple centers of private wealth, each with its own interests and each competing with the others, can serve as checks on one another and on the state. Douglass North and Barry Weingast tell exactly this story about England after the Glorious Revolution: Parliament could credibly constrain the Crown after 1688 precisely because the merchant and landed classes had accumulated enough private wealth and organizational capacity to make constitutional constraints stick. Montesquieu made a version of the same argument about the nobility as a check on monarchical power; Tocqueville made another about voluntary associations as bulwarks against democratic despotism. The classical liberal tradition has long understood that a pluralistic civil society with many competing centers of wealth is among the most durable defenses against centralized tyranny.
But the protective function depends on a condition that is easy to miss: wealth must remain distributed across genuinely competing factions rather than coordinated into a dominant one. Once the centers of wealth coordinate—through trade associations, shared political investment, revolving-door networks, or simply convergent class interests—the pluralistic check becomes a unified faction, and the unified faction captures the state. That is precisely the transition Acemoglu and Robinson’s model captures: the elite’s ability to solve the collective action problem of political investment is what drives the captured-democracy result. The problem isn’t wealth inequality as such. It’s wealth coordinated into a politically dominant faction.
Neither Buchanan nor Ostrom fully solves the problem Acemoglu and Robinson identify. Buchanan’s generality principle depends on whether genuine generality can be enforced in a polity where the wealthy have strong incentives to carve out exceptions. Ostrom’s polycentric arrangements are vulnerable to capture at the meta-level: the formal recognition and protection of local self-governance is itself something concentrated power can undermine. But both represent serious classical liberal engagement with the underlying problem, and both point toward responses that go well beyond simply shrinking the state.
***
What, then, should we do about the immediate problem of wealth concentrations already large enough to exercise significant de facto political power? Here the honest answer is that no single policy lever is adequate, and Acemoglu and Robinson’s framework helps explain why. Wealth concentrations large enough to purchase political influence cannot be undone by a single institutional intervention, because any single intervention will itself be subject to the de facto power it is trying to constrain. The appropriate response is therefore plural rather than singular, and ongoing rather than decisive.
One natural direction is tax reform aimed at dynastic wealth. I’ve argued at length that an annual wealth tax of the sort championed by Gabriel Zucman is poorly matched to its stated goals, largely because of valuation and incentive problems that worsen as the tax grows. But more modest reforms within the income tax base are promising: treating death as a realization event and repealing step-up in basis would bring decades of accrued capital gains into the tax base at a natural accounting point, and a minimum tax on accrued gains for the ultra-wealthy would target extraordinary returns as they arise. My colleague Miranda Perry Fleischer has shown that closing the step-up loophole is “a plausible second-best solution” for those concerned about concentration. And Fleischer has also made the case for a Rignano tax—one that taxes old money more heavily than new by taxing inherited wealth at progressively higher rates as it rolls through the generations. Even Robert Nozick, in The Examined Life, proposed something close: restructuring inheritance so that “taxes will subtract from the possessions people can bequeath the value of what they themselves have received through bequests,” such that “an inheritance could not cascade down the generations.”
But tax reform isn’t a complete answer, as a glance at the current American political landscape makes clear. The oligarchic clique around the present administration is largely self-made rather than inherited, and would not have been constrained by any plausible inheritance tax regime. Taxation can work on the next generation; it can’t solve the problem of this one. That’s a reason to think about tax reform alongside other institutional responses rather than as a replacement for them. The kind of market-structure reform Lindsey and Teles advocate—narrower intellectual property rights, lower barriers to occupational entry, less regulatory interference with housing supply, reduced subsidies for financial risk-taking—attacks the economic rents that fund political influence at their source. The generality principle Buchanan pointed toward suggests a discipline for evaluating institutional proposals: do they treat citizens generally, or do they invite the carving out of special interests? And Ostrom’s polycentric approach suggests that the long-run defense against capture lies less in any particular centralized reform than in the cultivation of multiple, overlapping institutions capable of absorbing shocks and routing around failures. No one of these approaches is sufficient on its own. Together, they begin to match the problem’s scale.
***
Classical liberals should take wealth inequality seriously. Not because the sufficiency argument is wrong—I still think it’s largely right as a matter of philosophy. And not because classical liberals need to sign on to the progressive policy agenda—the classical liberal toolkit is often the most effective route to the ends progressives most want to reach. They should take it seriously because their own deepest commitment demands it. If you believe, with Simons, that “no one may be trusted with much power,” you cannot avert your eyes from the most potent engine of power concentration operating in modern democracies. Wealth converts to political influence. Political influence converts to favorable rules. Favorable rules convert to more wealth.
What I have tried to show in this essay is that the classical liberal tradition has more resources for thinking about this cycle than its critics often suppose—and that the work of taking it seriously is only beginning. The framework Acemoglu and Robinson offer should reshape how classical liberals think about the relationship between wealth and power. Buchanan and Ostrom offer directions for institutional response that go well beyond shrinking the state. None of this settles the question of what is to be done. But that, perhaps, is the point. The question of what wealth does to power is not one any of us can afford to wave away—and it is certainly not one that any single policy lever can be trusted to settle.
